Ideas move in fronts. A new one lands in one field, takes it by storm, then crosses into the next, and the currents already there are swallowed or pushed to the edge. Above: how much English-language books talked about each movement, measured. Below: where each front landed, field by field, drawn like fronts on a weather map.
2026
The chart is wider than this screen: swipe it sideways to move through time. It follows the playhead while playing.
Front line: one movement crossing fields, teeth point the way it travelsRevival: an old current returning in a lighter shadeBeachhead: the work that landed it in that fieldEvent: a shock that hit several fields at once
Fronts
Tap a front to light it in both panels and list where it landed. Tap it again to clear.
How this was made
The top panel is measured. It uses Google Books Ngram (English corpus, smoothing 5), sampled every five years from 1840 to 2020. “One river” stacks each name’s share of the attention given to all 21 names, so the total is always 100%. “Crests” scales each name to its own highest share and orders them by when that crest came, so the fronts march down the page.
Cleaning: years before a movement existed are dropped where the word meant something else or came from misdated scans (Marxism before 1845, behaviorism before 1912, existentialism before 1930, cybernetics before 1945, systems theory before 1945, cognitive science before 1970, artificial intelligence before 1950, complexity theory before 1965, postmodernism before 1960, mindfulness before 1975, new materialism before 2000). Pragmatism is cut after 1950, when the word turns mostly to its everyday political sense. The early bump in “critical realism” is an older American movement of the 1920s with the same name. “Structuralism” also counts “post-structuralism”, which is why it seems to crest so late.
Books keep discussing a movement long after it stops winning. Positivism grows in the 20th century mostly as the thing being criticised. So the measured river shows attention, not allegiance.
The lower panel is mapped, not measured. Each landing is dated by a named work. The swell after landing, the peak and the fade are judgements, as are which fields count. A row’s height shows how much of that field’s attention the fronts held together; when they compete, each front gets its share.
Philosophy is split into its two schools, Continental and Anglo-American, because several fronts landed in one and never in the other.
Has anyone drawn this before?
John B. Sparks’s Histomap (1931) drew 4,000 years of world history as one river of civilisations whose widths show relative power.
Alfred Barr’s chart for Cubism and Abstract Art (MoMA, 1936) drew art movements as arrows feeding one another, the classic diagram of fronts in art.
Pitirim Sorokin’s Social and Cultural Dynamics (1937–41) counted philosophers’ systems of truth period by period over 2,500 years and found long cyclic swings, the nearest earlier attempt to measure waves of ideas.
Derek de Solla Price’s “Networks of Scientific Papers” (1965) named the research front: the small, fast-moving edge of recent work that a field cites most.
Randall Collins’s The Sociology of Philosophies (1998) mapped thinkers across world civilisations by generation and found a law of small numbers: only a few rival schools hold the attention space at once.
ThemeRiver (Havre, Hetzler and Nowell, 2000) is the river form used in the top panel: themes as currents in one river, widths as strength.
Rosvall and Bergstrom’s alluvial diagrams (2010) measured fields of science splitting and merging from citation data, the closest thing to streams entering a river.